图像

Based on the summarization of the present research, the future system frame of the blind digital image forensics is proposed. According to the natural characteristics of digital image, the principles of constructing the standard testing database of sample images are also put forward. Moreover, the extensible topology structure is set up and the size of the database needed theoretically is calculated statistically.

在总结前人工作的基础上，提出了数字图像盲取证的系统框架；利用数字图像自然属性特征，根据数字图像取证的测试需求，提出了取证测试专用的标准图像样本数据库建库准则，建立了可拓展的图像库拓扑结构，并利用统计学方法计算了所需的建库规模。

An image fusion scheme based on the wavelet transform and an image compression algorithm based on zero-tree coding are proposed, aiming at the demand of the Intelligent Transportation System. The algorithm is used to fuse visible image and infrared image with wavelet transform and then to compress the fused image by fully employing the importance degree of the subband coefficients and visual characteristic of mankind. So it enhances the integrality and oneness of the traffic information availably and also improves the quality of the system greatly.

针对智能交通系统信息采集的需要，提出基于小波分析的图像融合和零树编码的小波图像压缩算法，利用小波变换将红外和可见光图像进行融合，利用各子带小波系数对图像恢复的重要程度和人眼的视觉特性对融合图像进行压缩，算法的实际应用提高了所采集的交通信息的完整性、统一性和容错性，并改善了智能交通系统的服务质量。

At the aspect of effect assessment , the image quality assessment assume the knowledge of a reference image and blind image quality assessment are introduced according to the uncomplimentary information of IR images ,which benefit to evaluate the effect of the enhancement method of the IR image under different condition .

在对红外图像增强处理的效果评价方面，针对所掌握红外图像信息的完整性，分别就有参照图像的评价和盲图像的评价引入了不同的评价机制，从而为不同情况下的红外图像增强方法的有效性评估奠定了基础。

A great deal of experimental results are presented, proving that the digital watermark generated by the algorithm is unperceivable and shows great robustness to the common image processing like low-pass filtering, median filtering, Wiener filtering, nosing, cropping, rotation and JPEG loss compression, etc.

大量的实验结果表明，用本算法产生的数字水印能保证数字图像的不可见性，并且对常见图像处理均具有很好的鲁棒性，主要表现在低通滤波、中值滤波、维纳滤波、图像加噪，图像裁剪、图像旋转和有损JPEG压缩处理。

The phylogeny and applied prospect of MES are summarized in the paper. The principle of this system is analyzed from the point of view of acoustics. The surveying and filtering technology of auxiliary parameters, sounding data and the theory of image form, backscatter strength data processing, are researched deeply in the paper. The optimal model of velocity of sound is given. For sound ray tracing, some methods of correction of sound ray curve are developed. At the same time, an ideal model of sound ray tracing is proved and given. The GPS technology is applied for the determination of tide, attitude of vessel. Moreover, some methods are given. As an important topic, the systematic error is discussed in the paper. Some filtering methods are researched in this paper. Based on the current methods of image processing, the procedure and methods are tested and given that adapt to sonar image processing.

本文研究的工作是在总结多波束声纳测深系统的发展历史、应用前景的基础上，从声学角度分析了多波束系统的工作原理；对多波束辅助参数的测定和滤波方法、测深数据的滤波方法、多波束声纳图像的形成机理、图像处理以及最终应用等作了比较深入的研究；得出了适合于多波束精密深度计算的最优声速模型和声线改正模型；提出了 GPS 载波相位定位技术用于多波束辅助参数测量的思想、算法，改善了多波束系统的作业模式；系统地分析了深度数据的误差来源和类型，首次将系统误差作为影响测深精度的主要因素给以研究，并给出了削弱系统误差的方法，提高了深度数据的质量；对多波束声纳图像的形成、处理方法进行了细致的研究，给出了适合多波束图像处理的算法和流程，并从理论和实践上对声纳图像的应用进行了比较全面的研究。

In the studying of image segmentation, six crucial technologies havebeen developed successfully. 1 In one dimension histogram of original DRimage of luggage, there remain maximum-value wave crest havingapproximately stationary position, and the first minimum to the left of thewave crest performs excellently to be used to get rid of background of DRimage. 2 Open operation, one type of math morphology, is the right idea toerase luggage disturbance in DR image and at the same time, can be retaininteresting region appropriately. 3 After open operation, there are only a few minimums on image histogram, one of which can be to segment image furtherand achieves satisfying result. 4 Method of histogram analyzing and that oferosion performs well in the re-segmentation of conglutinative region in DRimage. 5 Method of image distance switch can ensure Luggage CT sectionautomatically obtained to have rich suspicious region information. 6 Todecompose structure element of math morphology is a way to save a great dealrunning time for automatically discerning interesting regions algorithm,apparently improving implementation efficiency.

在对图像可疑区域分割研究中，主要开发成功以下几项关键技术：1发现原始箱包DR图像的一维直方图中，均具有位置相对稳定的最大值波峰，以波峰左边第一个极小值为分割阈值，去除DR图像背景，效果非常好。2运用数学形态学中的开运算能够非常有效地去除箱子的干扰，同时能够比较好地保留可疑区域。3发现开运算后的DR图像直方图只有若干个极小值点的特点，其中用最大值波峰左边的第二个极小值点进一步分割DR图像，能够更有效地提取箱包中的可疑区域。4运用直方图分析法和腐蚀法，对粘连区域实行再分割，效果较好。5运用距离变换法，能确保自动设定的箱包CT截面包含丰富的可疑区域信息。6运用形态学结构元素分解技术，大幅度减少可疑区域的自动判别算法运行时间，提高算法软件执行效率。

This paper describes a new algorithm of image enhancement of impulse noise based on PCNN time matrix and rough sets indiscernibility relation. For noisy gray images, the results of computer simulations show that the image noise can be reduced efficiently and the image becomes clear and the target details of the image can be enhanced better by using this algorithm. In addition, the effects of this algorithm are better than usual image enhancement algorithm.

基于粗集理论中的不可分辨关系的等价概念及PCNN赋时矩阵，提出了一种含有脉冲噪声的图像增强算法，计算机仿真结果表明，对被噪声污染的灰度图像，用基于PCNN与粗集理论相结合的图像增强算法处理后，图像噪声得到有效抑制，画面更清晰，图像中目标的细节也得到了很好的增强，其结果优于常规方法。

This paper presents a new method of digital image sharing.Firstly,construct a reversible matrix,decompose a secret image into several scrambling and encryption images at one time by matrix operation,then camouflage them into several megascopic and meaningful carrier images to store or transfer.This algorithm combines the scrambling technique,encryption technique,information hiding and sharing technique,but with much simpler algorithm,less computation time,smaller pixel expansion,better recovery effect.The results show that the algorithm has application value in the field of secret image storage and transmission.

提出一种数字图像分存方法，首先构造可逆矩阵，并通过矩阵运算直接将秘密图像分解成几幅等大小的置乱加密图像，然后分别将它们伪装在几幅经过放大且有意义的载体图像中，实现了图像分存，文中将置乱加密、信息隐藏、分存技术三者有机结合，整个算法实现简单，运算量小，数据膨胀率低，安全性高，秘图恢复效果较好，大量实验也验证了算法的有效性。

The Fresnel lens in the optical receptor spotlights the headlamp beams to form their image at the 1-meter distance between the optical receptor and headlamps. CMOS camera as an image sensor, takes the photos of their images, which are captured by VG-M201, The application of some digital image processing techniques to these images, such as edge-picking up、 least-square fitting-curve algorithm, is to calculate the offset position/angle of optical axis of the high beams and the nodical offset of the unsymmetrical facular light and shade cut-off line of the low beams.

系统的图像采集工作是利用受光箱中的菲涅耳透镜在1米的测量距离上聚合前照灯灯光，由CMOS摄像头作为图像传感器件，照入灯光远光光斑和非对称近光图像，由嵌入式处理器对捕获的图像运用边缘提取、最小二乘法拟合曲线等数字图像处理技术计算出远光主光轴偏移角度以及非对称近光光斑明暗截止线拐点的偏移，实现了前照灯近光非对称光斑的自动测量。

The basic principle is that basic geometry shape of openings in the photo is revised; the image can't be distortion by use of projection reconstruction; the corrected image is exerted by open operation and close operation of gray-image; the shape of openings is gained lastly. The perimeter borderline of openings shape is extracted. The quantity of overbreak as well as underbreak is gained by comparing to extracted perimeter and designed perimeter. The same method is applied to other opening sections. Evaluation of overbreak as well as underbreak in the openings can be easily solved in the way of quantitative.

其基本原理是：对野外摄取的洞室的基本剖面形状进行几何校正，然后应用图像学中的投影重建方法使图像不失真，对修正后的图像实行灰度图像的开操作和闭操作运算，得到洞室的基本形状，然后用图像学中的边界提取技术提取洞室的周边界，与洞室的设计形状对比分析，得到洞室在该断面上的超欠挖方量，然后对其它断面采用同样的技术就可评价某一段或整个洞室的超欠挖数量，从定量的角度解决了工程中的超欠挖评价问题。

The split between the two groups can hardly be papered over.

这两个团体间的分歧难以掩饰。

This approach not only encourages a greater number of responses, but minimizes the likelihood of stale groupthink.

这种做法不仅鼓励了更多的反应，而且减少跟风的可能性。